This week's book giveaway is in the Android forum. We're giving away four copies of Head First Android and have Dawn & David Griffiths on-line! See this thread for details.

why we can't avoid hash collision in practice ? Hashmap use array of array to store value, the reason is hash key value collision. Then why we can't avoid it ? for me, I didn't see how difficult it is, we just use a primary key like in the db ...

Well, sometimes you can avoid it, and sometimes you can't. What if the key is a String? How many different possible Strings are there that might go into a Map? Hint: it's a really, really big number. The hashCode() must be an int however. How many different possible ints are there? Hint: Integer.MAX_VALUE is a big number, but much much smaller than the previous big number. Therefore, out of all possible Strings, some Strings must have the same hashCode() - no matter how you choose to calculate hashCode(). This leads to the possibility of hash collisions.

Let's have a look at String objects that consist of exactly 10 letters of the alphabet (A-Z). How many different strings are possible? The answer is: 26^10 = 141.167.095.653.376 different strings.

The hash code of an object is a 32-bit integer. So, how many different hash codes are possible? 2^32 = 4.294.967.296 different hash codes.

You see that there are many more possible strings than there are hash codes. Therefore, there must be different strings that have the same hash code.